AI outreach studio for B2B sales
Replylark drafts the whole sequence — opener, bump, break-up — and scores every email for deliverability while you edit, before anything sends.
Log in as demo@replylark.com · demo1234 — two audiences and their sequences are already threaded.
→ audience: SaaS founders, 11–50 seats
✎ drafting 4-step sequence…
✓ opener scored 94 — no spam triggers
✓ bump & break-up queued, day 3 / day 8
Subject (28 chars)
reporting day at {{company}}
Hi {{firstName}},
Most agency founders I talk to lose two full days a month to client reporting — and the clients skim it in ninety seconds…
4 emails
per sequence
94/100
demo opener’s deliverability
day 8
the polite break-up sends
2 min
brief to full thread
Capabilities
Cold email fails at the copy stage or the deliverability stage. Replylark holds both in one editor, so what leaves is worth sending and lands in the inbox.
Pain intro, value-add, case study, breakup — the structure that gets replies, written from your audience profile in one click.
Every email is scored against seven inbox heuristics — spam triggers, length, links, personalization — live as you edit.
"Act now", "risk-free", "guarantee" — the phrases that route you to Promotions get flagged the moment you type them.
ICP, pain, offer, proof — captured once, folded into every sequence so nothing starts from a blank page.
{{firstName}} and {{company}} tokens are checked into every email and paste cleanly into Instantly, Smartlead or Lemlist.
Works instantly with the built-in draft engine; connect an OpenAI or Anthropic key for fully bespoke sequences.
The workflow
Step 01
Who they are, what hurts, what you offer, one proof point. Once.
Step 02
New quarter, new feature, old pain — one line steers the whole sequence.
Step 03
Edit until every email scores green, copy into your sender, launch.
Plans
For your first campaign
$19/mo
For active pipelines
$49/mo
For teams & agencies
$99/mo
Spam-trigger phrasing, link and image weight, subject length, personalisation depth and send cadence — scored per email as you edit, with the fix named next to each deduction.
No — and that's deliberate. You send through your existing tool (Instantly, Smartlead, Lemlist, or plain Gmail) where your domain warmup and sending reputation live. Replylark writes and scores; connectors are on the documented roadmap.
Seven heuristics per email: subject length and casing, body word count, personalization tokens, spam-trigger phrases, link count, and whether it ends with a clear next step.
No — the built-in draft engine writes complete sequences instantly from your audience profile. Connect an OpenAI or Anthropic key in Settings for fully bespoke writing.
The opposite: it refuses the phrases that get you filtered and flags them if you add them back. Cold outreach still requires consent-law compliance on your end — the terms cover what's on you.
The demo workspace has a live sequence mid-edit — open it, watch the score move as you type, then draft your own.
Watch a sequence build →Log in as demo@replylark.com · demo1234 — two audiences and their sequences are already threaded.